Research map / Vision-language and multimodal / Robot action and navigation
Robo-ABC: Affordance Generalization Beyond Categories via Semantic Correspondence for Robot Manipulation
Yuanchen Ju, Kaizhe Hu, Guowei Zhang, Gu Zhang, Mingrun Jiang, Huazhe Xu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use retrieval from affordance memory plus diffusion-feature correspondence instead of end-to-end affordance models | Robo-ABC · AGD20K hold + door/drawer | SR (%): 29.1 -> 60.7 | Improved |
| use Robo-ABC contact points to select AnyGrasp grasps instead of VRB points | Franka real robot grasping · Real-world grasping (7 categories) | success rate (%): 68.6% -> 85.7% | Improved |
| use DIFT diffusion features instead of DINO-ViT for correspondence (ablation) | Robo-ABC · AGD20K hold + door/drawer | SR (%): 36.4 -> 60.7 | Improved |
| use DIFT instead of LDM-SC for correspondence (ablation) | Robo-ABC · AGD20K hold + door/drawer | SR (%): 52.0 -> 60.7 | Improved |
| add LPIPS to CLIP-B32 retrieval (ablation) | Robo-ABC (DIFT) · AGD20K hold + door/drawer | SR (%): 60.7 -> 50.2 | Hurt |
| use CLIP-B16 instead of CLIP-B32 retriever (ablation) | Robo-ABC (DIFT) · AGD20K hold + door/drawer | SR (%): 60.7 -> 60.0 | No clear effect |
| grow memory from 18 to 51 object categories (ablation) | Robo-ABC (DIFT) · AGD20K hold + door/drawer | SR (%): 56.4 -> 60.7 | Improved |
| grow memory from 18 to 51 categories (ablation) | DINO-ViT correspondence · AGD20K hold + door/drawer | SR (%): 39.3 -> 36.4 | Hurt |
| retrieve top-5 images instead of top-1 (ablation) | Robo-ABC (DIFT) · AGD20K hold + door/drawer | SR (%): 54.5 -> 60.7 | Improved |
| retrieve top-5 instead of top-3 images (ablation) | SD-DINO correspondence · AGD20K hold + door/drawer | SR (%): 59.3 -> 58.2 | Hurt |
Closest papers by meaning
- AffordDP: Generalizable Diffusion Policy with Transferable Affordance (CVPR 2025 · similarity 0.91)
- IAAO: Interactive Affordance Learning for Articulated Objects in 3D Environments (CVPR 2025 · similarity 0.89)
- Grounding 3D Object Affordance with Language Instructions, Visual Observations and Interactions (CVPR 2025 · similarity 0.89)
- VidBot: Learning Generalizable 3D Actions from In-the-Wild 2D Human Videos for Zero-Shot Robotic Manipulation (CVPR 2025 · similarity 0.89)
- GREAT: Geometry-Intention Collaborative Inference for Open-Vocabulary 3D Object Affordance Grounding (CVPR 2025 · similarity 0.89)
- SeqAfford: Sequential 3D Affordance Reasoning via Multimodal Large Language Model (CVPR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
